4.6 The CREATE_MODEL2 Procedure
The CREATE_MODEL2
procedure of the
DBMS_DATA_MINING
package is a procedure for defining model settings to
build a model.
By using the CREATE_MODEL2
procedure, the user does not need to
create transient database objects. The model can use configuration settings and
user-specified transformations. In the CREATE_MODEL2
procedure, the
input is a table or a view and if such an object is not already present, the user
must create it.
DBMS_DATA_MINING.CREATE_MODEL2 ( model_name IN VARCHAR2, mining_function IN VARCHAR2, data_query IN CLOB, set_list IN SETTING_LIST, case_id_column_name IN VARCHAR2 DEFAULT NULL, target_column_name IN VARCHAR2 DEFAULT NULL, xform_list IN TRANSFORM_LIST DEFAULT NULL);
The data_query
parameter species a query which provides training
data for building the model. The set_list
parameter specifies the
SETTING_LIST
. SETTING_LIST
is a table of CLOB
index by VARCHAR2(30)
; Where the index is the setting name and the
CLOB is the setting value for that name. The rest of the parameters are covered in
the CREATE_MODEL
procedure.
You can also rename the model using the RENAME_MODEL
procedure of
the DBMS_DATA_MINING
package. The procedure changes the value of
the machine learning model specified against MODEL_NAME
with
another name that you specify.
The following CREATE_MODEL2
procedure builds a classification model
using SVM algorithm. The following example mining_data_build_v data set to arrive at
likelihood of customers opting the affinity card program.
.
DECLARE v_setlist DBMS_DATA_MINING.SETTING_LIST; BEGIN v_setlist('PREP_AUTO') := 'ON'; v_setlist('ALGO_NAME') := 'ALGO_SUPPORT_VECTOR_MACHINES'; v_setlist('SVMS_KERNEL_FUNCTION') := 'SVMS_LINEAR'; DBMS_DATA_MINING.CREATE_MODEL2( MODEL_NAME => 'SVM_MODEL', MINING_FUNCTION => 'CLASSIFICATION', DATA_QUERY => 'select * from mining_data_build_v', SET_LIST => v_setlist, CASE_ID_COLUMN_NAME => 'CUST_ID, TARGET_COLUMN_NAME => 'AFFINITY_CARD'); END;